Skip to content

Linearity Analyzer

Linearity Analyzer

Overview

This tool precisely measures the "proportional relationship (linearity)" between input and output as the signal level changes. It is used to verify problems specific to digital equipment (DAC/ADC), such as "loud sounds are output correctly, but errors occur with very quiet sounds (bit truncation or buried in noise)," as well as the dynamic range of analog circuits.

Meaning of Key Indicators

This tool measures the following values:

  • Linearity Error
    • Represents how much the volume deviates compared to the reference gain (usually at maximum volume).
    • In ideal equipment, if the input is -100dB, the output attenuates by exactly -100dB, and the error is 0dB.
  • Linear Range
    • The minimum signal level where the error is within a certain range (e.g., ±0.5dB) and is not buried in the noise floor.
    • The lower this value (e.g., -120dBFS), the more capable the equipment is of accurately reproducing minute sounds.
  • Slope
    • Ideally 0 (flat), but changes if compression or similar effects are present.
  • Hysteresis Width
    • The difference in measurement values between the forward path (decreasing level) and the reverse path (increasing level). In electronic equipment, a value closer to 0 is better. If there is a significant difference, it may indicate thermal effects or delayed response due to the internal state of the device.

Operation

Running a Measurement

  1. Determine the measurement conditions in Sweep Settings.
  2. Press the Start Sweep button to begin measurement.
  3. The tool automatically sweeps the specified range (e.g., -5dBFS to -120dBFS) and plots the results on the graph.
    • If Enable Hysteresis Sweep is enabled, it performs forward (large to small) and reverse (small to large) measurements sequentially.

Interpreting Results

  • Linearity Error (Deviation) Graph
    • The horizontal axis is the input level, and the vertical axis is the error (dB).
    • The closer the line sticks to "0dB" in the center, the better.
    • The graph will start to fluctuate as it moves to the left (lower level region), which is due to the influence of noise.
  • Noise Floor Region (Gray Area)
    • Indicates the region where the SNR (Signal-To-Noise Ratio) is poor and the measured values are unreliable.

Settings

Sweep Settings

Sets the range and resolution of the measurement.

  • Frequency: Measurement frequency (usually 1kHz).
  • Start Level: Start level of the sweep (loud sound). Usually set to around -5 dBFS to avoid clipping.
  • End Level: End level of the sweep (quiet sound). Set to the limit value you want to verify (e.g., -120 dBFS).
  • Steps: Number of measurement points. More steps mean more detail but take longer.
  • SNR Limit: Threshold for noise floor detection. If the signal is not louder than the noise by at least this value (e.g., 10dB), the data is considered "unreliable."
  • Averaging: The number of measurements per point. Increasing the count reduces the influence of noise and stabilizes measurements at low levels, but increases the total measurement time.

I/O Routing

  • Output: Channel to output the measurement signal.
  • Input: Channel to receive the measurement signal.
  • Enable Hysteresis Sweep: When checked, the tool automatically performs a reverse sweep (small to large) after the forward sweep (large to small) to verify the difference between the two directions.

Display

  • Unit: Switches the horizontal axis unit of the graph.
    • dBFS: Digital Full Scale reference.
    • dBV: Voltage reference (valid when calibration settings are active).

Plot Controls

  • Y-Axis Zoom: Changes the vertical axis zoom level of the graph (Linearity Error).
    • Auto: Automatically adjusts to fit the data.
    • 0.1 dB - 20.0 dB: Fixes the view to the specified range (±).
    • Tolerance Lines: Green dashed lines indicating a tolerance of ±1.0 dB are displayed on the graph.

Usage Examples

Checking DAC Low-Level Reproduction Capability

Investigate whether your audio interface or DAC has 16-bit precision (approx. -96dB) or true 24-bit precision (around -144dB) capability.

  1. Connect in loopback (connect Output and Input directly with a cable).
  2. Set End Level to -140 dBFS and sweep.
  3. Check the graph to see how far it keeps 0dB error.
    • If it fluctuates around -90dB, it is 16-bit class performance.
    • If it holds up to -110dB to -120dB, it is very high performance.

Observing Hysteresis in Dual ADC Range Switching

Some recent recorders and high-end interfaces that support 32-bit float recording are equipped with two ADCs, and they achieve a vast dynamic range by internally switching gain depending on the input level. To prevent frequent clicking noises, this switching point is sometimes provided with "hysteresis" (processing that changes the threshold for switching between the forward and reverse paths).

  1. Start Level: -3 dBFS (Start here in this example).
  2. End Level: -9 dBFS (Lower to here).
  3. Check Enable Hysteresis Sweep.
  4. When the measurement is executed, a "gap" may appear in the gain or linearity error graph between the forward path (-3 -> -9) and the reverse path (-9 -> -3).
    • If this is observed, it means the device is performing gain switching in the analog or digital stage, and you have successfully visualized that behavior.